RE: FW: [NVDA] #809: Support for longdesc in web browsers

Hi Silvia,

Silvia Pfeiffer wrote:
>
> It is good to see more support for long description
> type features in screenreaders.

But let's be specific here Silvia - support for @longdesc. Not some
un-named, un-specified mechanism, but rather an existing, valid HTML4/XHTML1
attribute. If nothing it strengthens the case that if HTML5 is supposed to
be backward compatible then it is time to recognize @longdesc as a valid
attribute in the markup language.

>
> That actually supports the argument made in [1] that we
> need a screen-reader only long description attribute (such
> as the proposed @aria-describedat) and should mark
> @longdesc deprecated for HTML5 and obsolete when
> @aria-describedat has arrived.


I personally (and I think others) are not fans of making crystal ball
predictions on what may happen sometime in the future. 

There is no justifiable reason to make @longdesc deprecated now or into the
immediate future, as even if/when a future "better technique" emerges it
will take a number of years to see wide-spread adoption and take-up by end
users (just ask all those IE 8 users out there...). The Accessibility Task
Force are moving forward with proceeding on Chaals' Extension Specification
(http://dvcs.w3.org/hg/html-proposals/raw-file/b63325998cc1/longdesc1/longde
sc.html), and it is my understanding that once it reaches FPWD status that
this will remove the ambiguity around @longdesc's fate in HTML5.

With due respect to Mr. Mazzoni, @longdesc is NOT a feature for "web
developers", @longdesc is a feature for web consumers. I read his opinion
when you first posted it, and again just now, and I respectfully suggest
that he's missing a critical point in this debate, and is looking at this
issue strictly from a developer's perspective.

With commercial content creators (for example Pearson Publishing) and US
Federally funded initiatives (NCAM's Diagram Project -
http://ncam.wgbh.org/experience_learn/educational_media/diagram) producing
effective and valuable long text descriptions delivered via @longdesc today,
and, as evidenced by yesterday's NVDA announcement, more and more practical
support in screen reading tools, it strikes me as counter-productive to be
saying in HTML5 that we're gonna toss this baby out as soon as the new
bathwater is ready, which will be soon, but we don't know when. I would
prefer to see a much more orderly transition with some real overlap, and I
think that until we have a proposed solution (likely aria-describedat),
along with some implementation examples and experience with the new
attribute, that it is way too early to be discussing deprecation. 

I think that instead, we should be looking to encourage those browsers and
OS stacks that currently do not provide good support to @longdesc (as JF
glances towards Cupertino) to give it a second look - as NVDA did - and seek
instead to deliver a better user-experience for all users moving forward
today. Why wait? I have long argued that it is not the markup attribute that
needs work, it's the support from the user-agents. The screen readers are
(for the most part, and increasingly) doing their part, now we need the
browsers to do something useful here too.

JF

Received on Thursday, 8 November 2012 22:58:01 UTC